Station of the Week

These two stations, Dhahran on the Gulf coast of Saudi Arabia and Bahrain (an island in the Gulf), are only 31 miles apart. But their mutual trends make little sense at all. Bahrain shows a long-term warming trend of about 1°C, but Dhahran shows a `U' shaped trend, warm in the mid-century period, cooler during the 1970s and then warming again during the 1980s and 1990s., with the warmest period in the 1950s.

1999 was the warmest year at Bahrain, but 1961 was the warmest at Dhahran. The sharp cooling in the 1970s at Dhahran does not show up at Bahrain. 1933 was the coolest year at Bahrain, while 1972 was the coolest at Dhahran.

What are we to make of such contradictory trends from two stations so close together, and both influenced by the sea?

Clearly, they can't both be right, and it would be scientific folly to merely average them together. Unlike many station pairs shown on `Station of the Week' which validate each other with mutually consistent trends, these two invalidate each other since their trends are in no way mutually consistent.

Station data from large regions of the world are equally contradictory which is one reason why the satellite record is the preferred method of determining the temperature trend of the world as a whole.

Commentary by Dr Tim Ball:

You're correct to question the inconsistencies between the two stations Dhahrain and Bahrain. I have studied a heat island, Winnipeg Manitoba, and written on the issue. Winnipeg was a good study because it is located on a large flat plain and there are no urban areas within 500 km. There is also a distinct difference in seasons.

I have also compared weather data from juxtaposed weather records to determine the amount of relative homogeneity. It is critical in long term temperature studies to identify changes due to local effect from
regional and hemispheric trends. The problem is made more difficult when the stations straddle distinct natural climate boundaries such as along the Polar Front in both hemispheres.

Heat island studies are correctly blamed for false readings of local temperatures but ike so much that is done these days it is a simplistic approach. Initially we examined global temperatures then someone pointed out the difference in temperatures between the two hemispheres. If you separate summer and winter temperatures in most records you often have different trends. For example, summer temperatures usually show little warming and a distinct correlation with solar patterns, while winter temperatures reflect the warming since approximately 1680 AD. We also find differences in trends when we look at day and night temperatures. I have not heard clear explanation of any of these differences. A most likely candidate is water vapour which varies with air temperature.

Finally, heat island effects vary depending on the time of day, the month and the general trend in weather in the hemisphere or region. For example, the impact of the Winnipeg heat island was directly a result of the wind direction. The weather station at Winnipeg is to the northwest of the city. We found that the area of maximum impact of the heat island, that is the highest temperature distortions varied day and night, from month to month and season to season. The pattern of waves that form in the Jet Stream determines the general pattern of winds at the surface in the middle latitudes. Patterns could prevail for quite long periods thus changing the amount of influence on temperature readings. Chagnon studied such influences on precipitation with the classic LaPorte weather anomaly of patterns downwind of Chicago.

Dhahrain and Bahrain are subtropical stations so the effects of the Rossby waves is likely not an explanation of wind pattern changes. However, wind patterns change in these latitudes as El Niño/La Niña reversals indicate. I would like to know where the weather stations are at the two places relative to the influencing heat island. I suspect this might explain much of the difference in the patterns.

Incidentally, it is not just urban heat islands that can cause problems with temperature readings, as you pointed out a couple of years ago with your comments about Australian and Tasmanian weather sites.

Large heat islands are not always the problem. Edmonton, Alberta weather is recorded west of the city well outside severe heat island effects. However, the station is 100 metres south of a major four lane highway. In winter the effect of the heat from the heavy traffic and the blacktop road surface would unquestionably alter the readings at this station. The black top surface is important because it means distinctly different albedo in a snow covered landscape.

Commentary by Miceal O'Ronain

The temperature exchange between Bahrain & Dhahran are seasonally dependant. Dhahran is warmer during the Spring and Summer, while Bahrain is warmer during the Fall and Winter. Attached is a graph which shows this. Note that for the time period of 1950 to 1960 Dhahram is warmer and 1969 to 1983 when Bahrain was warmer. From the season graphics it would appear the Dhahran is the area which is undergoing the major changes in temperature. Bahrain is relatively constant in temperature for each season.